Update README.md

README.md

@@ -12,6 +12,8 @@ library_name: transformers
 <span style="font-family: default; font-size: 1.5em;">FastCuRL-1.5B-Preview</span>
 </div>
 
+## FastCuRL Overview
+
 We release **FastCuRL-1.5B-Preview**, a slow-thinking reasoning model that achieves 43.1% accuracy on the AIME 2024 benchmark! We adapt a novel curriculum-guided iterative lengthening reinforcement learning to the distilled 1.5B model and observe continuous performance improvement as training steps increase. To better reproduce our work and advance research progress, we open-source our code, model, and data.
 
 Code: https://github.com/nick7nlp/FastCuRL
@@ -31,6 +33,12 @@ We report Pass@1 accuracy averaged over 16 samples for each problem.
 | DeepScaleR-1.5B-Preview | 43.1 | 87.8 | 73.6 | 30.2 | 50.0 | 57.0 |
 | <strong>FastCuRL-1.5B-Preview</strong> | <strong>43.1</strong> | <strong>88.0</strong> | <strong>74.2</strong> | 31.6 | <strong>50.4</strong> | <strong>57.5</strong> |
 
+## Training Data
+Following DeepScaleR, our training dataset consists of 40,315 unique problem-answer pairs compiled from:
+- AIME problems (1984-2023)
+- AMC problems (before 2023)
+- Omni-MATH dataset
+- Still dataset
 
 ## Acknowledgements